DB: Consumer De-leveraging to End in Two Years
By Walter Kurtz, Sober Look
Deutsche Bank’s latest research on US consumer leverage suggests that the deleveraging process may have another couple of years to run. They determined the long-term trend line based on consumer debt to GDP ratio from 1953 to 2003, thus excluding the bubble years. Then they compared the current leverage to this line and looked at the rate of convergence. The intersection with the trend line is expected to take place in a year.
To be on the conservative side, they also assumed another year for the leverage ratio to overshoot to the downside and/or the nominal GDP to grow at slower rate than 4%. Note that the 4.0% nominal US GDP growth projection is not unrealistic (that’s roughly where it has been recently). They verified this result by repeating the exercise using debt to disposable income measure rather than debt to GDP.
DB: – … extrapolating the current trend further, household debt to GDP would intersect its long-term trend line by Q3 2013. It may be that households overcompensate, or nominal growth is less than 4.0% so we are conservatively estimating two years for the end of the deleveraging process. The story is broadly similar when comparing household debt to gross disposable income. This ratio has declined from a peak of 123% in Q1 2009 to 107% as of Q1 2012—the lowest level since Q3 2003 (106%). This tells us that consumers are nearing the end of the deleveraging process.
 |
| Source: DB |
18 Comments
This is really sad. We’re in such s–t condition as a country that we’ll be happy when this debt-to-GDP ratio returns to a point on an upward-sloping trend line. Decade after decade, this ratio increases and our situation gets uglier and uglier, and yet that’s okay because there’s a trendline?
I agree. First, there’s the assumption that growth from 1953-2003 was entirely organic (ignoring numerous large wars etc. over that period), then the picking of 2003 as the arbitrary “this is the correct level of debt and we’re almost back to it now” date used as “confirmation”. By that year the Chinese had been printing like maniacs for decades, the US money supply had been entirely disconnected from gold for decades — right around when median real net income started stagnating, the financial sector had been creating as much new money as possible since deregulation began decades prior — especially via OTC derivatives and shadow banking, and we’d already just had colossal stock market crashes in 1987 and 2000. Interest rates have arguably been much too low and bank regulation much too lax for a very long time (the “Great Moderation”), and yet it is declared that the credit bubble “started” in 2003, and thus everything will be fine again in a couple of years.
It’s amazing how an analysis can be done without a thought to the reality behind the numbers and trend lines. Just because the growth rate of the debt changed significantly in 2003 doesn’t mean that’s when the “bubble started”, it could equally mean that’s when the bubble’s growth accelerated.
(I’m not saying all growth prior to 2003 was a bubble — i.e., misallocated resources — but I also don’t think it’s wise to say none of it was. Is a multi-decade global credit expansion driven by unchecked financial innovation really that hard to imagine?)
Finally, after the Great Depression, IIRC, the household debt/GDP ratio fell below 50% before it started to increase again — and this time we took our debt levels even higher before we hit the wall. Thus there is no justification for picking 106% as the magic number representing a floor on the debt/GDP ratio aside from “that’s where an arbitrary line on the chart happens to pass through”, IMO.
It seems to me that it would be more sensible to find the last time at which the debt/GDP ratio was stable, indicating all the new credit being issued was actually being backed by productive things and the economy was thus structured in a sustainable configuration.
The chart suggests that the US consumer, after deleveraging for two more years, starts to increase their debtload.
DB is simply talking its own book. There’re in the business of lending, remember ?
LOL The US consumer is financially safe and sound soon!
The DB analysis starts as follows:
“extrapolating the current trend further”
This is the problem, it’s not just DB but almost all these simplicistic analyses are based on the assumption that there is a linear trend immutable in time with more or less frequent under/over shooting. This is badly wrong. Social/economical systems are changing their composition over time, feedback is at work. A statistician/biologist will tell you that many natural systems are not stationary and not ergodic that is concepts like mean, revert to the mean, standard deviation etc… can be misleading. But I was favourably impressed by most of the comments above: they are from people that understand this fact quite well.
Indeed, you usually can’t use mean or standard deviation when you’re talking about asset price changes, because they follow a power-law distribution, not a Gaussian one. Thus most of modern finance theory is nonsense. The problem with statistics is that certain techniques can only be used with certain distributions of data, and you’ll go badly wrong if you aren’t aware of that (such as calling a price movement a “six-sigma event” when there’s no such thing as “sigma” for a power-law distribution, or taking the mean of a bimodal distribution).
Oh finally ! Someone who knows about the fractal nature of the markets. Mandelbrot studied it 40 years ago but his conclusions are terrible for the finance industry. It’s almost impossibile to build a trading system on it; on the contrary it’s possibile to build a risk model but with the effect that risks must be priced much more and in that case the derivatives industry is dead. But robustness is not a concept mankind is ready for. It will learn it soon or will not survive the next 50 years.
I would like to add to your statement that even the linear trend is an unsustainable budget path for the US-consumer, since there is no Fed/US-consumer relationship like the relationship for the US government debt.
I would extrapolate the Debt to GDP back down to the 50% level area to be healthy for the economy. If the trend continues down that would put us toward the end of the decade. How would one factor in government debt effects on all of this? Paying back government debt will extend out even further as more private debt is reverted to pay down public debt will it not?
When I extrapolate the graph back to 50% I get to mid 2015.
no Gov debt does not need to be paid back, It is merely “spending” issued by the printing presses, and therefore does not need to be paid back. the great “debt myth” has been thoroughly debunked on this site… the difference btwn a currency issuer like the US Gov and a currency user like a household, or like a US State, or like Greece, is lost on 99% of mainstream econononsense
wow, so the sustainable level of debt to gdp increases over time linearly? so in 100 years the sustainable ratio will be something like 160%? I really wonder somtimes…
So, they claim that in two years we will see again sustainable GDP growth of 4% and consumer debt growth inline with previous trend from somewhere around 75%?
Does anybody here believes it? Sorry, but to me it sounds like a fairy tale…. But again, maybe I am nuts….
On the serious side, I would greatly like to see how they “verified” this with debt to disposable income calculations. It surprises me Cullen did not post that part. What are assumptions for income growth, taxes, inflation, and interest expense on debt outstanding? I have a bad feeling the Grimm brothers from DB assumed that we will be Japan with China GDP growth and Singapore taxes….
the regression line makes sense unless we have a new trend coming up.
With such an impressive back room research team as this, I think it might be a great time to short DB
Good Idea!
We focus maybe to much on the debt and not the GDP side. If we dramatically increase growth (and quit spending so much) we can dramatically change the ratio. Much debate on how to do that but won’t happen with the current smidiots (Smart, idiots) in Washington.